Although high dose three dimensionally planned radiation therapy improves local control and, possibly, survival for patients with focal intrahepatic cancers, many patients have diffuse disease. Amifostine has been shown in randomized trials to protect the parotid gland, lung, and esophagus from radiation. We propose to optimize the use of amifostine as a radiation protector of normal liver, which will permit the safe delivery of higher doses of radiation for patients with both focal and diffuse disease. In Specific Aim 1 we will conduct a phase I trial of dose escalating radiation therapy with systemic amifostine for patients with diffuse intrahepatic cancer. Our preclinical data demonstrate that systemic administration of amifostine produces more WR-1065 in the normal liver than in intrahepatic tumor, and that this leads to radioprotection of the liver (and not the tumor). Therefore, we hypothesize that systemic amifostine will permit meaningful selective protection of the normal liver, permitting radiation dose escalation. In Specific Aim 2 we will carry out preclinical studies to optimize selectivity (Aim 2A) and estimate the appropriate dose of regional amifostine (Aim 2B). Our preliminary data demonstrate that both systemic and portal venous amifostine protect normal liver without protecting tumor, and that portal venous amifostine produces significantly higher normal tissue/tumor ratios of the active metabolite WR-1065 than systemic amifostine. We hypothesize that regional amifostine will be superior to systemic amifostine in producing a higher liver to tumor ratio of WR-1065, causing greater selective protection of normal liver compared to tumor. In Specific Aim 3 we will conduct a phase I trial of dose escalating radiation therapy with regional amifostine for patients with diffuse intrahepatic cancer. We hypothesize that regional administration of amifostine will permit greater (or equal) protection of normal liver than is possible by systemic administration, with equal (or less) systemic toxicity. Our preliminary data, research team, and record for translating preclinical findings to the clinic suggest that this proposal is likely to improve the outcome of treatment of patients with diffuse intrahepatic cancer.